Skip to contentSkip to conversation

iommu: Refactoring domain allocation interface

PATCH Superseded Lu Baolu 6 participants 50 messages Jun 10, 2024 – Jul 9, 2024

Subsystems (29)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Lu Baolu

author maintainer
#0 lore raw

v3 cover iommu: Refactoring domain allocation interface

The IOMMU subsystem has undergone some changes, including the removal of iommu_ops from the bus structure. Consequently, the existing domain allocation interface, which relies on a bus type argument, is no longer relevant:

struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)

This series is designed to refactor the use of this interface. It proposes a new interface iommu_paging_domain_alloc() to replace iommu_domain_alloc(). This interface is for allocating iommu paging domains for kernel drivers. It takes a device pointer as a parameter, which better reflects the current design of the IOMMU subsystem.

The majority of device drivers currently using iommu_domain_alloc() do so to allocate a domain for a specific device and then attach that domain to the device. These cases can be straightforwardly migrated to the new interface.

The drm/tegra driver is a bit different in that the device pointer passed to the helper, which allocates the iommu domain, is not the one that will be used for the kernel DMA API. Move the existing logic in iommu_domain_alloc() into the driver to ensure it works as intended.

Now that all consumers of iommu_domain_alloc() have switched to the new interface, we can finally remove iommu_domain_alloc(). This removal paves the way for the IOMMU subsystem to support multiple iommu drivers. Additionally, the individual iommu driver implementation for domain allocation could also be simplified, as there will always be a valid device pointer passed along the path.

The whole series is also available on GitHub:
https://github.com/LuBaolu/intel-iommu/commits/iommu-domain-allocation-refactor-v3

Change log:

v3:
- Remove the proposed iommu_user_domain_alloc() interface.

v2: https://lore.kernel.org/linux-iommu/[email protected]/
- Drop the vt-d patches which implement paging domain support from this series. I will post them in a separate series later.
- Convert all drivers that call iommu_domain_alloc() to use the new interface and remove iommu_domain_alloc() from the tree.
- For the drm/msm driver, make the code compatible with the no-IOMMU
case.
- Various cleanups and refinements.

v1: https://lore.kernel.org/linux-iommu/[email protected]/

Lu Baolu (20):
iommu: Add iommu_paging_domain_alloc() interface
iommufd: Use iommu_paging_domain_alloc() vfio/type1: Use iommu_paging_domain_alloc()
vhost-vdpa: Use iommu_paging_domain_alloc() drm/msm: Use iommu_paging_domain_alloc() drm/nouveau/tegra: Use iommu_paging_domain_alloc()
gpu: host1x: Use iommu_paging_domain_alloc()
media: nvidia: tegra: Use iommu_paging_domain_alloc()
media: venus: firmware: Use iommu_paging_domain_alloc()
wifi: ath10k: Use iommu_paging_domain_alloc()
wifi: ath11k: Use iommu_paging_domain_alloc()
remoteproc: Use iommu_paging_domain_alloc() soc/fsl/qbman: Use iommu_paging_domain_alloc() RDMA/usnic: Use iommu_paging_domain_alloc() iommu/vt-d: Add helper to allocate paging domain
ARM: dma-mapping: Use iommu_paging_domain_alloc() drm/rockchip: Use iommu_paging_domain_alloc() drm/tegra: Remove call to iommu_domain_alloc()
iommu: Remove iommu_present()
iommu: Remove iommu_domain_alloc()

Robin Murphy (1):
ARM: dma-mapping: Pass device to arm_iommu_create_mapping()

 include/linux/iommu.h                         | 12 +--
 arch/arm/include/asm/dma-iommu.h              |  2 +-
 arch/arm/mm/dma-mapping.c                     | 12 +--
 drivers/gpu/drm/exynos/exynos_drm_dma.c       |  2 +-
 drivers/gpu/drm/msm/msm_iommu.c               |  7 +-
 .../drm/nouveau/nvkm/engine/device/tegra.c    |  4 +-
 drivers/gpu/drm/rockchip/rockchip_drm_drv.c   | 10 ++-
 drivers/gpu/drm/tegra/drm.c                   | 34 ++++++--
 drivers/gpu/host1x/dev.c                      |  7 +-
 drivers/infiniband/hw/usnic/usnic_uiom.c      |  6 +-
 drivers/iommu/intel/iommu.c                   | 87 +++++++++++++++++--
 drivers/iommu/iommu.c                         | 63 +++-----------
 drivers/iommu/iommufd/hw_pagetable.c          |  7 +-
 drivers/iommu/ipmmu-vmsa.c                    |  3 +-
 drivers/iommu/mtk_iommu_v1.c                  |  3 +-
 .../media/platform/nvidia/tegra-vde/iommu.c   |  7 +-
 drivers/media/platform/qcom/venus/firmware.c  |  6 +-
 drivers/media/platform/ti/omap3isp/isp.c      |  2 +-
 drivers/net/wireless/ath/ath10k/snoc.c        |  6 +-
 drivers/net/wireless/ath/ath11k/ahb.c         |  6 +-
 drivers/remoteproc/remoteproc_core.c          |  6 +-
 drivers/soc/fsl/qbman/qman_portal.c           |  5 +-
 drivers/vfio/vfio_iommu_type1.c               |  7 +-
 drivers/vhost/vdpa.c                          | 14 ++-
 24 files changed, 183 insertions(+), 135 deletions(-)
Signature · 1 line
2.34.1
author maintainer
#1 lore raw

PATCH v3 1/21 iommu: Add iommu_paging_domain_alloc() interface

Commit <17de3f5fdd35> ("iommu: Retire bus ops") removes iommu ops from bus. The iommu subsystem no longer relies on bus for operations. So the bus parameter in iommu_domain_alloc() is no longer relevant.

Add a new interface named iommu_paging_domain_alloc(), which explicitly indicates the allocation of a paging domain for DMA managed by a kernel driver. The new interface takes a device pointer as its parameter, that better aligns with the current iommu subsystem.

---
 include/linux/iommu.h |  6 ++++++
 drivers/iommu/iommu.c | 20 ++++++++++++++++++++
 2 files changed, 26 insertions(+)
include/linux/iommu.h +6 -0
index 7bc8dff7cf6d..46f1348f1f0a 100644
@@ -780,6 +780,7 @@ extern bool iommu_present(const struct bus_type *bus);
 extern bool device_iommu_capable(struct device *dev, enum iommu_cap cap);
 extern bool iommu_group_has_isolated_msi(struct iommu_group *group);
 extern struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus);
+struct iommu_domain *iommu_paging_domain_alloc(struct device *dev);
 extern void iommu_domain_free(struct iommu_domain *domain);
 extern int iommu_attach_device(struct iommu_domain *domain,
 			       struct device *dev);
@@ -1086,6 +1087,11 @@ static inline struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus
 	return NULL;
 }
 
+static inline struct iommu_domain *iommu_paging_domain_alloc(struct device *dev)
+{
+	return ERR_PTR(-ENODEV);
+}
+
 static inline void iommu_domain_free(struct iommu_domain *domain)
 {
 }
drivers/iommu/iommu.c +20 -0
index 9df7cc75c1bc..e03c71a34347 100644
@@ -2016,6 +2016,10 @@ static int __iommu_domain_alloc_dev(struct device *dev, void *data)
 	return 0;
 }
 
+/*
+ * The iommu ops in bus has been retired. Do not use this interface in
+ * new drivers.
+ */
 struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus)
 {
 	const struct iommu_ops *ops = NULL;
@@ -2032,6 +2036,22 @@ struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus)
 }
 EXPORT_SYMBOL_GPL(iommu_domain_alloc);
 
+/**
+ * iommu_paging_domain_alloc() - Allocate a paging domain
+ * @dev: device for which the domain is allocated
+ *
+ * Allocate a paging domain which will be managed by a kernel driver. Return
+ * allocated domain if successful, or a ERR pointer for failure.
+ */
+struct iommu_domain *iommu_paging_domain_alloc(struct device *dev)
+{
+	if (!dev_has_iommu(dev))
+		return ERR_PTR(-ENODEV);
+
+	return __iommu_domain_alloc(dev_iommu_ops(dev), dev, IOMMU_DOMAIN_UNMANAGED);
+}
+EXPORT_SYMBOL_GPL(iommu_paging_domain_alloc);
+
 void iommu_domain_free(struct iommu_domain *domain)
 {
 	if (domain->type == IOMMU_DOMAIN_SVA)
Signature · 1 line
2.34.1
Reviewed-by on v3 1/21
#2 lore raw

Re: [PATCH v3 01/21] iommu: Add iommu_paging_domain_alloc() interface

Quoted 9 lines · Lu Baolu

Commit <17de3f5fdd35> ("iommu: Retire bus ops") removes iommu ops from bus. The iommu subsystem no longer relies on bus for operations. So the bus parameter in iommu_domain_alloc() is no longer relevant.

Add a new interface named iommu_paging_domain_alloc(), which explicitly indicates the allocation of a paging domain for DMA managed by a kernel driver. The new interface takes a device pointer as its parameter, that better aligns with the current iommu subsystem.

Quoted 5 lines
---
 include/linux/iommu.h |  6 ++++++
 drivers/iommu/iommu.c | 20 ++++++++++++++++++++
 2 files changed, 26 insertions(+)

Jason

Reviewed-by on v3 1/21
#3 lore raw

Re: [PATCH v3 01/21] iommu: Add iommu_paging_domain_alloc() interface

In reply to Lu Baolu
Quoted 10 lines · Lu Baolu

Commit <17de3f5fdd35> ("iommu: Retire bus ops") removes iommu ops from bus. The iommu subsystem no longer relies on bus for operations. So the bus parameter in iommu_domain_alloc() is no longer relevant.

Add a new interface named iommu_paging_domain_alloc(), which explicitly indicates the allocation of a paging domain for DMA managed by a kernel driver. The new interface takes a device pointer as its parameter, that better aligns with the current iommu subsystem.

-Vasant

Quoted 67 lines
---
 include/linux/iommu.h |  6 ++++++
 drivers/iommu/iommu.c | 20 ++++++++++++++++++++
 2 files changed, 26 insertions(+)
include/linux/iommu.h +6 -0
index 7bc8dff7cf6d..46f1348f1f0a 100644
@@ -780,6 +780,7 @@ extern bool iommu_present(const struct bus_type *bus);
 extern bool device_iommu_capable(struct device *dev, enum iommu_cap cap);
 extern bool iommu_group_has_isolated_msi(struct iommu_group *group);
 extern struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus);
+struct iommu_domain *iommu_paging_domain_alloc(struct device *dev);
 extern void iommu_domain_free(struct iommu_domain *domain);
 extern int iommu_attach_device(struct iommu_domain *domain,
 			       struct device *dev);
@@ -1086,6 +1087,11 @@ static inline struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus
 	return NULL;
 }
 
+static inline struct iommu_domain *iommu_paging_domain_alloc(struct device *dev)
+{
+	return ERR_PTR(-ENODEV);
+}
+
 static inline void iommu_domain_free(struct iommu_domain *domain)
 {
 }
drivers/iommu/iommu.c +20 -0
index 9df7cc75c1bc..e03c71a34347 100644
@@ -2016,6 +2016,10 @@ static int __iommu_domain_alloc_dev(struct device *dev, void *data)
 	return 0;
 }
 
+/*
+ * The iommu ops in bus has been retired. Do not use this interface in
+ * new drivers.
+ */
 struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus)
 {
 	const struct iommu_ops *ops = NULL;
@@ -2032,6 +2036,22 @@ struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus)
 }
 EXPORT_SYMBOL_GPL(iommu_domain_alloc);
 
+/**
+ * iommu_paging_domain_alloc() - Allocate a paging domain
+ * @dev: device for which the domain is allocated
+ *
+ * Allocate a paging domain which will be managed by a kernel driver. Return
+ * allocated domain if successful, or a ERR pointer for failure.
+ */
+struct iommu_domain *iommu_paging_domain_alloc(struct device *dev)
+{
+	if (!dev_has_iommu(dev))
+		return ERR_PTR(-ENODEV);
+
+	return __iommu_domain_alloc(dev_iommu_ops(dev), dev, IOMMU_DOMAIN_UNMANAGED);
+}
+EXPORT_SYMBOL_GPL(iommu_paging_domain_alloc);
+
 void iommu_domain_free(struct iommu_domain *domain)
 {
 	if (domain->type == IOMMU_DOMAIN_SVA)
author maintainer
#4 lore raw

PATCH v3 2/21 iommufd: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

If the iommu driver doesn't implement its domain_alloc_user callback, iommufd_hwpt_paging_alloc() rolls back to allocate an iommu paging domain. Replace iommu_domain_alloc() with iommu_user_domain_alloc() to pass the device pointer along the path.

---
 drivers/iommu/iommufd/hw_pagetable.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/iommu/iommufd/hw_pagetable.c +4 -3
index 33d142f8057d..cbddfa3ca95a 100644
@@ -137,9 +137,10 @@ iommufd_hwpt_paging_alloc(struct iommufd_ctx *ictx, struct iommufd_ioas *ioas,
 		}
 		hwpt->domain->owner = ops;
 	} else {
-		hwpt->domain = iommu_domain_alloc(idev->dev->bus);
-		if (!hwpt->domain) {
-			rc = -ENOMEM;
+		hwpt->domain = iommu_paging_domain_alloc(idev->dev);
+		if (IS_ERR(hwpt->domain)) {
+			rc = PTR_ERR(hwpt->domain);
+			hwpt->domain = NULL;
 			goto out_abort;
 		}
 	}
Signature · 1 line
2.34.1
Reviewed-by on v3 2/21
#5 lore raw

Re: [PATCH v3 02/21] iommufd: Use iommu_paging_domain_alloc()

Quoted 9 lines · Lu Baolu

If the iommu driver doesn't implement its domain_alloc_user callback, iommufd_hwpt_paging_alloc() rolls back to allocate an iommu paging domain. Replace iommu_domain_alloc() with iommu_user_domain_alloc() to pass the device pointer along the path.

---
 drivers/iommu/iommufd/hw_pagetable.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

Jason

author maintainer
#6 lore raw

PATCH v3 3/21 vfio/type1: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

Replace iommu_domain_alloc() with iommu_paging_domain_alloc().

---
 drivers/vfio/vfio_iommu_type1.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/vfio/vfio_iommu_type1.c +4 -3
index 3a0218171cfa..0960699e7554 100644
@@ -2135,7 +2135,7 @@ static int vfio_iommu_domain_alloc(struct device *dev, void *data)
 {
 	struct iommu_domain **domain = data;
 
-	*domain = iommu_domain_alloc(dev->bus);
+	*domain = iommu_paging_domain_alloc(dev);
 	return 1; /* Don't iterate */
 }
 
@@ -2192,11 +2192,12 @@ static int vfio_iommu_type1_attach_group(void *iommu_data,
 	 * us a representative device for the IOMMU API call. We don't actually
 	 * want to iterate beyond the first device (if any).
 	 */
-	ret = -EIO;
 	iommu_group_for_each_dev(iommu_group, &domain->domain,
 				 vfio_iommu_domain_alloc);
-	if (!domain->domain)
+	if (IS_ERR(domain->domain)) {
+		ret = PTR_ERR(domain->domain);
 		goto out_free_domain;
+	}
 
 	if (iommu->nesting) {
 		ret = iommu_enable_nesting(domain->domain);
Signature · 1 line
2.34.1
author maintainer
#8 lore raw

PATCH v3 4/21 vhost-vdpa: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

Replace iommu_domain_alloc() with iommu_paging_domain_alloc().

---
 drivers/vhost/vdpa.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)
drivers/vhost/vdpa.c +6 -8
index 63a53680a85c..e31ec9ebc4ce 100644
@@ -1312,26 +1312,24 @@ static int vhost_vdpa_alloc_domain(struct vhost_vdpa *v)
 	struct vdpa_device *vdpa = v->vdpa;
 	const struct vdpa_config_ops *ops = vdpa->config;
 	struct device *dma_dev = vdpa_get_dma_dev(vdpa);
-	const struct bus_type *bus;
 	int ret;
 
 	/* Device want to do DMA by itself */
 	if (ops->set_map || ops->dma_map)
 		return 0;
 
-	bus = dma_dev->bus;
-	if (!bus)
-		return -EFAULT;
-
 	if (!device_iommu_capable(dma_dev, IOMMU_CAP_CACHE_COHERENCY)) {
 		dev_warn_once(&v->dev,
 			      "Failed to allocate domain, device is not IOMMU cache coherent capable\n");
 		return -ENOTSUPP;
 	}
 
-	v->domain = iommu_domain_alloc(bus);
-	if (!v->domain)
-		return -EIO;
+	v->domain = iommu_paging_domain_alloc(dma_dev);
+	if (IS_ERR(v->domain)) {
+		ret = PTR_ERR(v->domain);
+		v->domain = NULL;
+		return ret;
+	}
 
 	ret = iommu_attach_device(v->domain, dma_dev);
 	if (ret)
Signature · 1 line
2.34.1
maintainer Acked-by on v3 4/21
#10 lore raw

Re: [PATCH v3 04/21] vhost-vdpa: Use iommu_paging_domain_alloc()

In reply to Lu Baolu
Quoted 3 lines · Lu Baolu

Replace iommu_domain_alloc() with iommu_paging_domain_alloc().

I assume it's merged with the rest of the stuff, right?

Quoted 43 lines
---
 drivers/vhost/vdpa.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)
drivers/vhost/vdpa.c +6 -8
index 63a53680a85c..e31ec9ebc4ce 100644
@@ -1312,26 +1312,24 @@ static int vhost_vdpa_alloc_domain(struct vhost_vdpa *v)
 	struct vdpa_device *vdpa = v->vdpa;
 	const struct vdpa_config_ops *ops = vdpa->config;
 	struct device *dma_dev = vdpa_get_dma_dev(vdpa);
-	const struct bus_type *bus;
 	int ret;
 
 	/* Device want to do DMA by itself */
 	if (ops->set_map || ops->dma_map)
 		return 0;
 
-	bus = dma_dev->bus;
-	if (!bus)
-		return -EFAULT;
-
 	if (!device_iommu_capable(dma_dev, IOMMU_CAP_CACHE_COHERENCY)) {
 		dev_warn_once(&v->dev,
 			      "Failed to allocate domain, device is not IOMMU cache coherent capable\n");
 		return -ENOTSUPP;
 	}
 
-	v->domain = iommu_domain_alloc(bus);
-	if (!v->domain)
-		return -EIO;
+	v->domain = iommu_paging_domain_alloc(dma_dev);
+	if (IS_ERR(v->domain)) {
+		ret = PTR_ERR(v->domain);
+		v->domain = NULL;
+		return ret;
+	}
 
 	ret = iommu_attach_device(v->domain, dma_dev);
 	if (ret)
Signature · 1 line
2.34.1
maintainer
#11 lore raw

Re: [PATCH v3 04/21] vhost-vdpa: Use iommu_paging_domain_alloc()

Quoted 9 lines · Michael S. Tsirkin
Quoted 3 lines · Lu Baolu

Replace iommu_domain_alloc() with iommu_paging_domain_alloc().

I assume it's merged with the rest of the stuff, right?

Sure, I can grab this one in the iommu tree. It has to go along with the patch adding iommu_paging_domain_alloc() anyway.

Will

author maintainer
#12 lore raw

PATCH v3 5/21 drm/msm: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

The domain allocated in msm_iommu_new() is for the @dev. Replace iommu_domain_alloc() with iommu_paging_domain_alloc() to make it explicit.

---
 drivers/gpu/drm/msm/msm_iommu.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/msm/msm_iommu.c +5 -2
index d5512037c38b..2a94e82316f9 100644
@@ -407,10 +407,13 @@ struct msm_mmu *msm_iommu_new(struct device *dev, unsigned long quirks)
 	struct msm_iommu *iommu;
 	int ret;
 
-	domain = iommu_domain_alloc(dev->bus);
-	if (!domain)
+	if (!device_iommu_mapped(dev))
 		return NULL;
 
+	domain = iommu_paging_domain_alloc(dev);
+	if (IS_ERR(domain))
+		return ERR_CAST(domain);
+
 	iommu_set_pgtable_quirks(domain, quirks);
 
 	iommu = kzalloc(sizeof(*iommu), GFP_KERNEL);
Signature · 1 line
2.34.1
author maintainer
#13 lore raw

PATCH v3 6/21 drm/nouveau/tegra: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

In nvkm_device_tegra_probe_iommu(), a paging domain is allocated for @dev and attached to it on success. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c +2 -2
index 87caa4a72921..763c4c2925f9 100644
@@ -120,8 +120,8 @@ nvkm_device_tegra_probe_iommu(struct nvkm_device_tegra *tdev)
 	mutex_init(&tdev->iommu.mutex);
 
 	if (device_iommu_mapped(dev)) {
-		tdev->iommu.domain = iommu_domain_alloc(&platform_bus_type);
-		if (!tdev->iommu.domain)
+		tdev->iommu.domain = iommu_paging_domain_alloc(dev);
+		if (IS_ERR(tdev->iommu.domain))
 			goto error;
 
 		/*
Signature · 1 line
2.34.1
author maintainer
#14 lore raw

PATCH v3 7/21 gpu: host1x: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in host1x_iommu_attach() and is attached to host->dev. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/gpu/host1x/dev.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/gpu/host1x/dev.c +4 -3
index 3a0aaa68ac8d..f86a6b12f24a 100644
@@ -404,9 +404,10 @@ static struct iommu_domain *host1x_iommu_attach(struct host1x *host)
 		if (err < 0)
 			goto put_group;
 
-		host->domain = iommu_domain_alloc(&platform_bus_type);
-		if (!host->domain) {
-			err = -ENOMEM;
+		host->domain = iommu_paging_domain_alloc(host->dev);
+		if (IS_ERR(host->domain)) {
+			err = PTR_ERR(host->domain);
+			host->domain = NULL;
 			goto put_cache;
 		}
 
Signature · 1 line
2.34.1
Reviewed-by on v3 7/21
#15 lore raw

Re: [PATCH v3 07/21] gpu: host1x: Use iommu_paging_domain_alloc()

Quoted 7 lines · Lu Baolu

An iommu domain is allocated in host1x_iommu_attach() and is attached to host->dev. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/gpu/host1x/dev.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

Jason

author maintainer
#16 lore raw

PATCH v3 8/21 media: nvidia: tegra: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in tegra_vde_iommu_init() and is attached to vde->dev. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/media/platform/nvidia/tegra-vde/iommu.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/media/platform/nvidia/tegra-vde/iommu.c +4 -3
index 5521ed3e465f..b1d9d841d944 100644
@@ -78,9 +78,10 @@ int tegra_vde_iommu_init(struct tegra_vde *vde)
 		arm_iommu_release_mapping(mapping);
 	}
 #endif
-	vde->domain = iommu_domain_alloc(&platform_bus_type);
-	if (!vde->domain) {
-		err = -ENOMEM;
+	vde->domain = iommu_paging_domain_alloc(dev);
+	if (IS_ERR(vde->domain)) {
+		err = PTR_ERR(vde->domain);
+		vde->domain = NULL;
 		goto put_group;
 	}
 
Signature · 1 line
2.34.1
Reviewed-by on v3 8/21
#17 lore raw

Re: [PATCH v3 08/21] media: nvidia: tegra: Use iommu_paging_domain_alloc()

Quoted 7 lines · Lu Baolu

An iommu domain is allocated in tegra_vde_iommu_init() and is attached to vde->dev. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/media/platform/nvidia/tegra-vde/iommu.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 

Jason

author maintainer
#18 lore raw

PATCH v3 9/21 media: venus: firmware: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in venus_firmware_init() and is attached to core->fw.dev in the same function. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/media/platform/qcom/venus/firmware.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/media/platform/qcom/venus/firmware.c +3 -3
index fe7da2b30482..66a18830e66d 100644
@@ -316,10 +316,10 @@ int venus_firmware_init(struct venus_core *core)
 
 	core->fw.dev = &pdev->dev;
 
-	iommu_dom = iommu_domain_alloc(&platform_bus_type);
-	if (!iommu_dom) {
+	iommu_dom = iommu_paging_domain_alloc(core->fw.dev);
+	if (IS_ERR(iommu_dom)) {
 		dev_err(core->fw.dev, "Failed to allocate iommu domain\n");
-		ret = -ENOMEM;
+		ret = PTR_ERR(iommu_dom);
 		goto err_unregister;
 	}
 
Signature · 1 line
2.34.1
Reviewed-by on v3 9/21
#19 lore raw

Re: [PATCH v3 09/21] media: venus: firmware: Use iommu_paging_domain_alloc()

Quoted 8 lines · Lu Baolu

An iommu domain is allocated in venus_firmware_init() and is attached to core->fw.dev in the same function. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/media/platform/qcom/venus/firmware.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Jason

author maintainer
#20 lore raw

PATCH v3 10/21 wifi: ath10k: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in ath10k_fw_init() and is attached to ar_snoc->fw.dev in the same function. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/net/wireless/ath/ath10k/snoc.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/wireless/ath/ath10k/snoc.c +3 -3
index 8530550cf5df..0fe47d51013c 100644
@@ -1635,10 +1635,10 @@ static int ath10k_fw_init(struct ath10k *ar)
 
 	ar_snoc->fw.dev = &pdev->dev;
 
-	iommu_dom = iommu_domain_alloc(&platform_bus_type);
-	if (!iommu_dom) {
+	iommu_dom = iommu_paging_domain_alloc(ar_snoc->fw.dev);
+	if (IS_ERR(iommu_dom)) {
 		ath10k_err(ar, "failed to allocate iommu domain\n");
-		ret = -ENOMEM;
+		ret = PTR_ERR(iommu_dom);
 		goto err_unregister;
 	}
 
Signature · 1 line
2.34.1
Reviewed-by on v3 10/21
#21 lore raw

Re: [PATCH v3 10/21] wifi: ath10k: Use iommu_paging_domain_alloc()

Quoted 9 lines · Lu Baolu

An iommu domain is allocated in ath10k_fw_init() and is attached to ar_snoc->fw.dev in the same function. Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/net/wireless/ath/ath10k/snoc.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Jason

author maintainer
#22 lore raw

PATCH v3 11/21 wifi: ath11k: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in ath11k_ahb_fw_resources_init() and is attached to ab_ahb->fw.dev in the same function.

Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/net/wireless/ath/ath11k/ahb.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/wireless/ath/ath11k/ahb.c +3 -3
index ca0f17ddebba..a469647719f9 100644
@@ -1001,10 +1001,10 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab)
 
 	ab_ahb->fw.dev = &pdev->dev;
 
-	iommu_dom = iommu_domain_alloc(&platform_bus_type);
-	if (!iommu_dom) {
+	iommu_dom = iommu_paging_domain_alloc(ab_ahb->fw.dev);
+	if (IS_ERR(iommu_dom)) {
 		ath11k_err(ab, "failed to allocate iommu domain\n");
-		ret = -ENOMEM;
+		ret = PTR_ERR(iommu_dom);
 		goto err_unregister;
 	}
 
Signature · 1 line
2.34.1
Reviewed-by on v3 11/21
#23 lore raw

Re: [PATCH v3 11/21] wifi: ath11k: Use iommu_paging_domain_alloc()

Quoted 10 lines · Lu Baolu

An iommu domain is allocated in ath11k_ahb_fw_resources_init() and is attached to ab_ahb->fw.dev in the same function.

Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/net/wireless/ath/ath11k/ahb.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Jason

author maintainer
#24 lore raw

PATCH v3 12/21 remoteproc: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in rproc_enable_iommu() and is attached to rproc->dev.parent in the same function.

Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/remoteproc/remoteproc_core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/remoteproc/remoteproc_core.c +3 -3
index f276956f2c5c..eb66f78ec8b7 100644
@@ -109,10 +109,10 @@ static int rproc_enable_iommu(struct rproc *rproc)
 		return 0;
 	}
 
-	domain = iommu_domain_alloc(dev->bus);
-	if (!domain) {
+	domain = iommu_paging_domain_alloc(dev);
+	if (IS_ERR(domain)) {
 		dev_err(dev, "can't alloc iommu domain\n");
-		return -ENOMEM;
+		return PTR_ERR(domain);
 	}
 
 	iommu_set_fault_handler(domain, rproc_iommu_fault, rproc);
Signature · 1 line
2.34.1
Reviewed-by on v3 12/21
#25 lore raw

Re: [PATCH v3 12/21] remoteproc: Use iommu_paging_domain_alloc()

Quoted 9 lines · Lu Baolu

An iommu domain is allocated in rproc_enable_iommu() and is attached to rproc->dev.parent in the same function.

Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/remoteproc/remoteproc_core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Jason

author maintainer
#26 lore raw

PATCH v3 13/21 soc/fsl/qbman: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

An iommu domain is allocated in portal_set_cpu() and is attached to pcfg->dev in the same function.

Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/soc/fsl/qbman/qman_portal.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/soc/fsl/qbman/qman_portal.c +3 -2
index e23b60618c1a..456ef5d5c199 100644
@@ -48,9 +48,10 @@ static void portal_set_cpu(struct qm_portal_config *pcfg, int cpu)
 	struct device *dev = pcfg->dev;
 	int ret;
 
-	pcfg->iommu_domain = iommu_domain_alloc(&platform_bus_type);
-	if (!pcfg->iommu_domain) {
+	pcfg->iommu_domain = iommu_paging_domain_alloc(dev);
+	if (IS_ERR(pcfg->iommu_domain)) {
 		dev_err(dev, "%s(): iommu_domain_alloc() failed", __func__);
+		pcfg->iommu_domain = NULL;
 		goto no_iommu;
 	}
 	ret = fsl_pamu_configure_l1_stash(pcfg->iommu_domain, cpu);
Signature · 1 line
2.34.1
Reviewed-by on v3 13/21
#27 lore raw

Re: [PATCH v3 13/21] soc/fsl/qbman: Use iommu_paging_domain_alloc()

Quoted 9 lines · Lu Baolu

An iommu domain is allocated in portal_set_cpu() and is attached to pcfg->dev in the same function.

Use iommu_paging_domain_alloc() to make it explicit.

---
 drivers/soc/fsl/qbman/qman_portal.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Jason

author maintainer
#28 lore raw

PATCH v3 14/21 RDMA/usnic: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

usnic_uiom_alloc_pd() allocates a paging domain for a given device. In this case, iommu_domain_alloc(dev->bus) is equivalent to iommu_paging_domain_alloc(dev). Replace it as iommu_domain_alloc() has been deprecated.

---
 drivers/infiniband/hw/usnic/usnic_uiom.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/infiniband/hw/usnic/usnic_uiom.c +3 -3
index 84e0f41e7dfa..f948b76f984d 100644
@@ -443,11 +443,11 @@ struct usnic_uiom_pd *usnic_uiom_alloc_pd(struct device *dev)
 	if (!pd)
 		return ERR_PTR(-ENOMEM);
 
-	pd->domain = domain = iommu_domain_alloc(dev->bus);
-	if (!domain) {
+	pd->domain = domain = iommu_paging_domain_alloc(dev);
+	if (IS_ERR(domain)) {
 		usnic_err("Failed to allocate IOMMU domain");
 		kfree(pd);
-		return ERR_PTR(-ENOMEM);
+		return ERR_CAST(domain);
 	}
 
 	iommu_set_fault_handler(pd->domain, usnic_uiom_dma_fault, NULL);
Signature · 1 line
2.34.1
author maintainer
#29 lore raw

PATCH v3 15/21 iommu/vt-d: Add helper to allocate paging domain

In reply to Lu Baolu

The domain_alloc_user operation is currently implemented by allocating a paging domain using iommu_domain_alloc(). This is because it needs to fully initialize the domain before return. Add a helper to do this to avoid using iommu_domain_alloc().

---
 drivers/iommu/intel/iommu.c | 87 +++++++++++++++++++++++++++++++++----
 1 file changed, 78 insertions(+), 9 deletions(-)
drivers/iommu/intel/iommu.c +78 -9
index 2e9811bf2a4e..ccde5f5972e4 100644
@@ -3633,6 +3633,79 @@ static struct iommu_domain blocking_domain = {
 	}
 };
 
+static int iommu_superpage_capability(struct intel_iommu *iommu, bool first_stage)
+{
+	if (!intel_iommu_superpage)
+		return 0;
+
+	if (first_stage)
+		return cap_fl1gp_support(iommu->cap) ? 2 : 1;
+
+	return fls(cap_super_page_val(iommu->cap));
+}
+
+static struct dmar_domain *paging_domain_alloc(struct device *dev, bool first_stage)
+{
+	struct device_domain_info *info = dev_iommu_priv_get(dev);
+	struct intel_iommu *iommu = info->iommu;
+	struct dmar_domain *domain;
+	int addr_width;
+
+	domain = kzalloc(sizeof(*domain), GFP_KERNEL);
+	if (!domain)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&domain->devices);
+	INIT_LIST_HEAD(&domain->dev_pasids);
+	INIT_LIST_HEAD(&domain->cache_tags);
+	spin_lock_init(&domain->lock);
+	spin_lock_init(&domain->cache_lock);
+	xa_init(&domain->iommu_array);
+
+	domain->nid = dev_to_node(dev);
+	domain->has_iotlb_device = info->ats_enabled;
+	domain->use_first_level = first_stage;
+
+	/* calculate the address width */
+	addr_width = agaw_to_width(iommu->agaw);
+	if (addr_width > cap_mgaw(iommu->cap))
+		addr_width = cap_mgaw(iommu->cap);
+	domain->gaw = addr_width;
+	domain->agaw = iommu->agaw;
+	domain->max_addr = __DOMAIN_MAX_ADDR(addr_width);
+
+	/* iommu memory access coherency */
+	domain->iommu_coherency = iommu_paging_structure_coherency(iommu);
+
+	/* pagesize bitmap */
+	domain->domain.pgsize_bitmap = SZ_4K;
+	domain->iommu_superpage = iommu_superpage_capability(iommu, first_stage);
+	domain->domain.pgsize_bitmap |= domain_super_pgsize_bitmap(domain);
+
+	/*
+	 * IOVA aperture: First-level translation restricts the input-address
+	 * to a canonical address (i.e., address bits 63:N have the same value
+	 * as address bit [N-1], where N is 48-bits with 4-level paging and
+	 * 57-bits with 5-level paging). Hence, skip bit [N-1].
+	 */
+	domain->domain.geometry.force_aperture = true;
+	domain->domain.geometry.aperture_start = 0;
+	if (first_stage)
+		domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw - 1);
+	else
+		domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw);
+
+	/* always allocate the top pgd */
+	domain->pgd = iommu_alloc_page_node(domain->nid, GFP_KERNEL);
+	if (!domain->pgd) {
+		kfree(domain);
+		return ERR_PTR(-ENOMEM);
+	}
+	domain_flush_cache(domain, domain->pgd, PAGE_SIZE);
+
+	return domain;
+}
+
 static struct iommu_domain *intel_iommu_domain_alloc(unsigned type)
 {
 	struct dmar_domain *dmar_domain;
@@ -3695,15 +3768,11 @@ intel_iommu_domain_alloc_user(struct device *dev, u32 flags,
 	if (user_data || (dirty_tracking && !ssads_supported(iommu)))
 		return ERR_PTR(-EOPNOTSUPP);
 
-	/*
-	 * domain_alloc_user op needs to fully initialize a domain before
-	 * return, so uses iommu_domain_alloc() here for simple.
-	 */
-	domain = iommu_domain_alloc(dev->bus);
-	if (!domain)
-		return ERR_PTR(-ENOMEM);
-
-	dmar_domain = to_dmar_domain(domain);
+	/* Do not use first stage for user domain translation. */
+	dmar_domain = paging_domain_alloc(dev, false);
+	if (IS_ERR(dmar_domain))
+		return ERR_CAST(dmar_domain);
+	domain = &dmar_domain->domain;
 
 	if (nested_parent) {
 		dmar_domain->nested_parent = true;
Signature · 1 line
2.34.1
Reviewed-by on v3 15/21
#30 lore raw

Re: [PATCH v3 15/21] iommu/vt-d: Add helper to allocate paging domain

Quoted 9 lines · Lu Baolu

The domain_alloc_user operation is currently implemented by allocating a paging domain using iommu_domain_alloc(). This is because it needs to fully initialize the domain before return. Add a helper to do this to avoid using iommu_domain_alloc().

---
 drivers/iommu/intel/iommu.c | 87 +++++++++++++++++++++++++++++++++----
 1 file changed, 78 insertions(+), 9 deletions(-)

It seems Ok, but I have some small thoughts

Quoted 26 lines
drivers/iommu/intel/iommu.c +18 -0
index 2e9811bf2a4e..ccde5f5972e4 100644
@@ -3633,6 +3633,79 @@ static struct iommu_domain blocking_domain = {
 	}
 };
 
+static int iommu_superpage_capability(struct intel_iommu *iommu, bool first_stage)
+{
+	if (!intel_iommu_superpage)
+		return 0;
+
+	if (first_stage)
+		return cap_fl1gp_support(iommu->cap) ? 2 : 1;
+
+	return fls(cap_super_page_val(iommu->cap));
+}
+
+static struct dmar_domain *paging_domain_alloc(struct device *dev, bool first_stage)
+{
+	struct device_domain_info *info = dev_iommu_priv_get(dev);
+	struct intel_iommu *iommu = info->iommu;
+	struct dmar_domain *domain;
+	int addr_width;
+

drivers/iommu/intel/iommu.c static struct iommu_domain blocking_domain = { · line 3663

Quoted 10 lines
drivers/iommu/intel/iommu.c +10 -0
+	domain = kzalloc(sizeof(*domain), GFP_KERNEL);
+	if (!domain)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&domain->devices);
+	INIT_LIST_HEAD(&domain->dev_pasids);
+	INIT_LIST_HEAD(&domain->cache_tags);
+	spin_lock_init(&domain->lock);
+	spin_lock_init(&domain->cache_lock);
+	xa_init(&domain->iommu_array);

You should probably split more, with an 'alloc struct dmar_domain' function that can be used by SVA and others too.

Quoted 9 lines
drivers/iommu/intel/iommu.c +9 -0
+	domain->nid = dev_to_node(dev);
+	domain->has_iotlb_device = info->ats_enabled;
+	domain->use_first_level = first_stage;
+
+	/* calculate the address width */
+	addr_width = agaw_to_width(iommu->agaw);
+	if (addr_width > cap_mgaw(iommu->cap))
+		addr_width = cap_mgaw(iommu->cap);
+	domain->gaw = addr_width;

drivers/iommu/intel/iommu.c static struct iommu_domain blocking_domain = { · line 3683

Quoted 10 lines
drivers/iommu/intel/iommu.c +10 -0
+	domain->agaw = iommu->agaw;
+	domain->max_addr = __DOMAIN_MAX_ADDR(addr_width);
+
+	/* iommu memory access coherency */
+	domain->iommu_coherency = iommu_paging_structure_coherency(iommu);
+
+	/* pagesize bitmap */
+	domain->domain.pgsize_bitmap = SZ_4K;
+	domain->iommu_superpage = iommu_superpage_capability(iommu, first_stage);
+	domain->domain.pgsize_bitmap |= domain_super_pgsize_bitmap(domain);

Then some of this stuff is really just paging only. Like SVA/identity/etc don't have pgszie and other things.

Jason

author maintainer
#31 lore raw

Re: [PATCH v3 15/21] iommu/vt-d: Add helper to allocate paging domain

Quoted 66 lines · Jason Gunthorpe
Quoted 9 lines · Lu Baolu

The domain_alloc_user operation is currently implemented by allocating a paging domain using iommu_domain_alloc(). This is because it needs to fully initialize the domain before return. Add a helper to do this to avoid using iommu_domain_alloc().

---
  drivers/iommu/intel/iommu.c | 87 +++++++++++++++++++++++++++++++++----
  1 file changed, 78 insertions(+), 9 deletions(-)

It seems Ok, but I have some small thoughts

Quoted 26 lines
drivers/iommu/intel/iommu.c +18 -0
index 2e9811bf2a4e..ccde5f5972e4 100644
@@ -3633,6 +3633,79 @@ static struct iommu_domain blocking_domain = {
  	}
  };
 
+static int iommu_superpage_capability(struct intel_iommu *iommu, bool first_stage)
+{
+	if (!intel_iommu_superpage)
+		return 0;
+
+	if (first_stage)
+		return cap_fl1gp_support(iommu->cap) ? 2 : 1;
+
+	return fls(cap_super_page_val(iommu->cap));
+}
+
+static struct dmar_domain *paging_domain_alloc(struct device *dev, bool first_stage)
+{
+	struct device_domain_info *info = dev_iommu_priv_get(dev);
+	struct intel_iommu *iommu = info->iommu;
+	struct dmar_domain *domain;
+	int addr_width;
+

drivers/iommu/intel/iommu.c static struct iommu_domain blocking_domain = { · line 3663

Quoted 10 lines
drivers/iommu/intel/iommu.c +10 -0
+	domain = kzalloc(sizeof(*domain), GFP_KERNEL);
+	if (!domain)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&domain->devices);
+	INIT_LIST_HEAD(&domain->dev_pasids);
+	INIT_LIST_HEAD(&domain->cache_tags);
+	spin_lock_init(&domain->lock);
+	spin_lock_init(&domain->cache_lock);
+	xa_init(&domain->iommu_array);

You should probably split more, with an 'alloc struct dmar_domain' function that can be used by SVA and others too.

Quoted 13 lines
drivers/iommu/intel/iommu.c +13 -0
+	domain->nid = dev_to_node(dev);
+	domain->has_iotlb_device = info->ats_enabled;
+	domain->use_first_level = first_stage;
+
+	/* calculate the address width */
+	addr_width = agaw_to_width(iommu->agaw);
+	if (addr_width > cap_mgaw(iommu->cap))
+		addr_width = cap_mgaw(iommu->cap);
+	domain->gaw = addr_width;
+	domain->agaw = iommu->agaw;
+	domain->max_addr = __DOMAIN_MAX_ADDR(addr_width);
+
+	/* iommu memory access coherency */
Quoted 8 lines

drivers/iommu/intel/iommu.c static struct iommu_domain blocking_domain = { · line 3683

Quoted 6 lines
drivers/iommu/intel/iommu.c +6 -0
+	domain->iommu_coherency = iommu_paging_structure_coherency(iommu);
+
+	/* pagesize bitmap */
+	domain->domain.pgsize_bitmap = SZ_4K;
+	domain->iommu_superpage = iommu_superpage_capability(iommu, first_stage);
+	domain->domain.pgsize_bitmap |= domain_super_pgsize_bitmap(domain);

Then some of this stuff is really just paging only. Like SVA/identity/etc don't have pgszie and other things.

Yeah, good suggestion. I could do this in this series if it needs a new version or in my next series which will add domain_alloc_paging for the intel driver.

Best regards,
baolu

#32 lore raw

Re: [PATCH v3 15/21] iommu/vt-d: Add helper to allocate paging domain

In reply to Lu Baolu
Quoted 99 lines · Lu Baolu

The domain_alloc_user operation is currently implemented by allocating a paging domain using iommu_domain_alloc(). This is because it needs to fully initialize the domain before return. Add a helper to do this to avoid using iommu_domain_alloc().

---
  drivers/iommu/intel/iommu.c | 87 +++++++++++++++++++++++++++++++++----
  1 file changed, 78 insertions(+), 9 deletions(-)
drivers/iommu/intel/iommu.c +73 -1
index 2e9811bf2a4e..ccde5f5972e4 100644
@@ -3633,6 +3633,79 @@ static struct iommu_domain blocking_domain = {
  	}
  };
 
+static int iommu_superpage_capability(struct intel_iommu *iommu, bool first_stage)
+{
+	if (!intel_iommu_superpage)
+		return 0;
+
+	if (first_stage)
+		return cap_fl1gp_support(iommu->cap) ? 2 : 1;
+
+	return fls(cap_super_page_val(iommu->cap));
+}
+
+static struct dmar_domain *paging_domain_alloc(struct device *dev, bool first_stage)
+{
+	struct device_domain_info *info = dev_iommu_priv_get(dev);
+	struct intel_iommu *iommu = info->iommu;
+	struct dmar_domain *domain;
+	int addr_width;
+
+	domain = kzalloc(sizeof(*domain), GFP_KERNEL);
+	if (!domain)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&domain->devices);
+	INIT_LIST_HEAD(&domain->dev_pasids);
+	INIT_LIST_HEAD(&domain->cache_tags);
+	spin_lock_init(&domain->lock);
+	spin_lock_init(&domain->cache_lock);
+	xa_init(&domain->iommu_array);
+
+	domain->nid = dev_to_node(dev);
+	domain->has_iotlb_device = info->ats_enabled;
+	domain->use_first_level = first_stage;
+
+	/* calculate the address width */
+	addr_width = agaw_to_width(iommu->agaw);
+	if (addr_width > cap_mgaw(iommu->cap))
+		addr_width = cap_mgaw(iommu->cap);
+	domain->gaw = addr_width;
+	domain->agaw = iommu->agaw;
+	domain->max_addr = __DOMAIN_MAX_ADDR(addr_width);
+
+	/* iommu memory access coherency */
+	domain->iommu_coherency = iommu_paging_structure_coherency(iommu);
+
+	/* pagesize bitmap */
+	domain->domain.pgsize_bitmap = SZ_4K;
+	domain->iommu_superpage = iommu_superpage_capability(iommu, first_stage);
+	domain->domain.pgsize_bitmap |= domain_super_pgsize_bitmap(domain);
+
+	/*
+	 * IOVA aperture: First-level translation restricts the input-address
+	 * to a canonical address (i.e., address bits 63:N have the same value
+	 * as address bit [N-1], where N is 48-bits with 4-level paging and
+	 * 57-bits with 5-level paging). Hence, skip bit [N-1].
+	 */
+	domain->domain.geometry.force_aperture = true;
+	domain->domain.geometry.aperture_start = 0;
+	if (first_stage)
+		domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw - 1);
+	else
+		domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw);
+
+	/* always allocate the top pgd */
+	domain->pgd = iommu_alloc_page_node(domain->nid, GFP_KERNEL);
+	if (!domain->pgd) {
+		kfree(domain);
+		return ERR_PTR(-ENOMEM);
+	}
+	domain_flush_cache(domain, domain->pgd, PAGE_SIZE);
+
+	return domain;
+}
+
  static struct iommu_domain *intel_iommu_domain_alloc(unsigned type)
  {
  	struct dmar_domain *dmar_domain;
@@ -3695,15 +3768,11 @@ intel_iommu_domain_alloc_user(struct device *dev, u32 flags,
  	if (user_data || (dirty_tracking && !ssads_supported(iommu)))
  		return ERR_PTR(-EOPNOTSUPP);
 
-	/*

drivers/iommu/intel/iommu.c intel_iommu_domain_alloc_user(struct device *dev, u32 flags, · line 3772

Quoted 10 lines
drivers/iommu/intel/iommu.c +2 -8
-	 * domain_alloc_user op needs to fully initialize a domain before
-	 * return, so uses iommu_domain_alloc() here for simple.
-	 */
-	domain = iommu_domain_alloc(dev->bus);
-	if (!domain)
-		return ERR_PTR(-ENOMEM);
-
-	dmar_domain = to_dmar_domain(domain);
+	/* Do not use first stage for user domain translation. */
+	dmar_domain = paging_domain_alloc(dev, false);

this is not an apple-to-apple replacement yet. You need to set the type, owner and domain->ops as well.

Quoted 6 lines
drivers/iommu/intel/iommu.c +3 -0
+	if (IS_ERR(dmar_domain))
+		return ERR_CAST(dmar_domain);
+	domain = &dmar_domain->domain;
 
  	if (nested_parent) {
  		dmar_domain->nested_parent = true;
Signature · 2 lines
Regards,
Yi Liu
author maintainer
#33 lore raw

Re: [PATCH v3 15/21] iommu/vt-d: Add helper to allocate paging domain

Quoted 111 lines · Yi Liu
Quoted 110 lines · Lu Baolu

The domain_alloc_user operation is currently implemented by allocating a paging domain using iommu_domain_alloc(). This is because it needs to fully initialize the domain before return. Add a helper to do this to avoid using iommu_domain_alloc().

---
  drivers/iommu/intel/iommu.c | 87 +++++++++++++++++++++++++++++++++----

  1 file changed, 78 insertions(+), 9 deletions(-)
drivers/iommu/intel/iommu.c +51 -0
index 2e9811bf2a4e..ccde5f5972e4 100644
@@ -3633,6 +3633,79 @@ static struct iommu_domain blocking_domain = {
       }
   };
+static int iommu_superpage_capability(struct intel_iommu *iommu, bool 
 first_stage)
+{
+    if (!intel_iommu_superpage)
+        return 0;
+
+    if (first_stage)
+        return cap_fl1gp_support(iommu->cap) ? 2 : 1;
+
+    return fls(cap_super_page_val(iommu->cap));
+}
+
+static struct dmar_domain *paging_domain_alloc(struct device *dev, 
 bool first_stage)
+{
+    struct device_domain_info *info = dev_iommu_priv_get(dev);
+    struct intel_iommu *iommu = info->iommu;
+    struct dmar_domain *domain;
+    int addr_width;
+
+    domain = kzalloc(sizeof(*domain), GFP_KERNEL);
+    if (!domain)
+        return ERR_PTR(-ENOMEM);
+
+    INIT_LIST_HEAD(&domain->devices);
+    INIT_LIST_HEAD(&domain->dev_pasids);
+    INIT_LIST_HEAD(&domain->cache_tags);
+    spin_lock_init(&domain->lock);
+    spin_lock_init(&domain->cache_lock);
+    xa_init(&domain->iommu_array);
+
+    domain->nid = dev_to_node(dev);
+    domain->has_iotlb_device = info->ats_enabled;
+    domain->use_first_level = first_stage;
+
+    /* calculate the address width */
+    addr_width = agaw_to_width(iommu->agaw);
+    if (addr_width > cap_mgaw(iommu->cap))
+        addr_width = cap_mgaw(iommu->cap);
+    domain->gaw = addr_width;
+    domain->agaw = iommu->agaw;
+    domain->max_addr = __DOMAIN_MAX_ADDR(addr_width);
+
+    /* iommu memory access coherency */
+    domain->iommu_coherency = iommu_paging_structure_coherency(iommu);
+
+    /* pagesize bitmap */
+    domain->domain.pgsize_bitmap = SZ_4K;
+    domain->iommu_superpage = iommu_superpage_capability(iommu, 
 first_stage);
+    domain->domain.pgsize_bitmap |= domain_super_pgsize_bitmap(domain);
+
+    /*
+     * IOVA aperture: First-level translation restricts the 

input-address
+     * to a canonical address (i.e., address bits 63:N have the same value
+     * as address bit [N-1], where N is 48-bits with 4-level paging and +     * 57-bits with 5-level paging). Hence, skip bit [N-1]. +     */
+    domain->domain.geometry.force_aperture = true;
+    domain->domain.geometry.aperture_start = 0;
+    if (first_stage)
+        domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw - 1);
+    else
+        domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw);

drivers/iommu/intel/iommu.c +12 -0
+
+    /* always allocate the top pgd */
+    domain->pgd = iommu_alloc_page_node(domain->nid, GFP_KERNEL);
+    if (!domain->pgd) {
+        kfree(domain);
+        return ERR_PTR(-ENOMEM);
+    }
+    domain_flush_cache(domain, domain->pgd, PAGE_SIZE);
+
+    return domain;
+}
+

  static struct iommu_domain *intel_iommu_domain_alloc(unsigned type) {
      struct dmar_domain *dmar_domain;

drivers/iommu/intel/iommu.c +0 -6
@@ -3695,15 +3768,11 @@ intel_iommu_domain_alloc_user(struct device 
 *dev, u32 flags,
       if (user_data || (dirty_tracking && !ssads_supported(iommu)))
           return ERR_PTR(-EOPNOTSUPP);
-    /*
-     * domain_alloc_user op needs to fully initialize a domain before
-     * return, so uses iommu_domain_alloc() here for simple.
-     */
-    domain = iommu_domain_alloc(dev->bus);
-    if (!domain)
Quoted 8 lines

drivers/iommu/intel/iommu.c intel_iommu_domain_alloc_user(struct device · line 3772

Quoted 5 lines
drivers/iommu/intel/iommu.c +2 -3
-        return ERR_PTR(-ENOMEM);
-
-    dmar_domain = to_dmar_domain(domain);
+    /* Do not use first stage for user domain translation. */
+    dmar_domain = paging_domain_alloc(dev, false);

this is not an apple-to-apple replacement yet. You need to set the type, owner and domain->ops as well.

Yes. You are right. I will add below lines.

drivers/iommu/intel/iommu.c +3 -0
index d51f840715e5..c5d9c2283977 100644
@@ -3732,6 +3732,9 @@ intel_iommu_domain_alloc_user(struct device *dev, u32 flags,
        if (IS_ERR(dmar_domain))
                return ERR_CAST(dmar_domain);
        domain = &dmar_domain->domain;
+       domain->type = IOMMU_DOMAIN_UNMANAGED;
+       domain->owner = &intel_iommu_ops;
+       domain->ops = intel_iommu_ops.default_domain_ops;
 
        if (nested_parent) {
                dmar_domain->nested_parent = true;

Best regards,
baolu

author maintainer
#34 lore raw

PATCH v3 16/21 ARM: dma-mapping: Pass device to arm_iommu_create_mapping()

In reply to Lu Baolu

From: Robin Murphy <[email protected]>

All users of ARM IOMMU mappings create them for a particular device, so change the interface to accept the device rather than forcing a vague indirection through a bus type. This prepares for making a similar change to iommu_domain_alloc() itself.

---
 arch/arm/include/asm/dma-iommu.h         | 2 +-
 arch/arm/mm/dma-mapping.c                | 8 ++++----
 drivers/gpu/drm/exynos/exynos_drm_dma.c  | 2 +-
 drivers/iommu/ipmmu-vmsa.c               | 3 +--
 drivers/iommu/mtk_iommu_v1.c             | 3 +--
 drivers/media/platform/ti/omap3isp/isp.c | 2 +-
 6 files changed, 9 insertions(+), 11 deletions(-)
arch/arm/include/asm/dma-iommu.h +1 -1
index 82ec1ccf1fee..2ce4c5683e6d 100644
@@ -24,7 +24,7 @@ struct dma_iommu_mapping {
 };
 
 struct dma_iommu_mapping *
-arm_iommu_create_mapping(const struct bus_type *bus, dma_addr_t base, u64 size);
+arm_iommu_create_mapping(struct device *dev, dma_addr_t base, u64 size);
 
 void arm_iommu_release_mapping(struct dma_iommu_mapping *mapping);
 
arch/arm/mm/dma-mapping.c +4 -4
index 5adf1769eee4..52f9c56cc3cb 100644
@@ -1532,7 +1532,7 @@ static const struct dma_map_ops iommu_ops = {
 
 /**
  * arm_iommu_create_mapping
- * @bus: pointer to the bus holding the client device (for IOMMU calls)
+ * @dev: pointer to the client device (for IOMMU calls)
  * @base: start address of the valid IO address space
  * @size: maximum size of the valid IO address space
  *
@@ -1544,7 +1544,7 @@ static const struct dma_map_ops iommu_ops = {
  * arm_iommu_attach_device function.
  */
 struct dma_iommu_mapping *
-arm_iommu_create_mapping(const struct bus_type *bus, dma_addr_t base, u64 size)
+arm_iommu_create_mapping(struct device *dev, dma_addr_t base, u64 size)
 {
 	unsigned int bits = size >> PAGE_SHIFT;
 	unsigned int bitmap_size = BITS_TO_LONGS(bits) * sizeof(long);
@@ -1585,7 +1585,7 @@ arm_iommu_create_mapping(const struct bus_type *bus, dma_addr_t base, u64 size)
 
 	spin_lock_init(&mapping->lock);
 
-	mapping->domain = iommu_domain_alloc(bus);
+	mapping->domain = iommu_domain_alloc(dev->bus);
 	if (!mapping->domain)
 		goto err4;
 
@@ -1718,7 +1718,7 @@ static void arm_setup_iommu_dma_ops(struct device *dev)
 		dma_base = dma_range_map_min(dev->dma_range_map);
 		size = dma_range_map_max(dev->dma_range_map) - dma_base;
 	}
-	mapping = arm_iommu_create_mapping(dev->bus, dma_base, size);
+	mapping = arm_iommu_create_mapping(dev, dma_base, size);
 	if (IS_ERR(mapping)) {
 		pr_warn("Failed to create %llu-byte IOMMU mapping for device %s\n",
 				size, dev_name(dev));
drivers/gpu/drm/exynos/exynos_drm_dma.c +1 -1
index e2c7373f20c6..6a6761935224 100644
@@ -110,7 +110,7 @@ int exynos_drm_register_dma(struct drm_device *drm, struct device *dev,
 		void *mapping = NULL;
 
 		if (IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU))
-			mapping = arm_iommu_create_mapping(&platform_bus_type,
+			mapping = arm_iommu_create_mapping(dev,
 				EXYNOS_DEV_ADDR_START, EXYNOS_DEV_ADDR_SIZE);
 		else if (IS_ENABLED(CONFIG_IOMMU_DMA))
 			mapping = iommu_get_domain_for_dev(priv->dma_dev);
drivers/iommu/ipmmu-vmsa.c +1 -2
index b657cc09605f..ff55b8c30712 100644
@@ -804,8 +804,7 @@ static int ipmmu_init_arm_mapping(struct device *dev)
 	if (!mmu->mapping) {
 		struct dma_iommu_mapping *mapping;
 
-		mapping = arm_iommu_create_mapping(&platform_bus_type,
-						   SZ_1G, SZ_2G);
+		mapping = arm_iommu_create_mapping(dev, SZ_1G, SZ_2G);
 		if (IS_ERR(mapping)) {
 			dev_err(mmu->dev, "failed to create ARM IOMMU mapping\n");
 			ret = PTR_ERR(mapping);
drivers/iommu/mtk_iommu_v1.c +1 -2
index d6e4002200bd..da61df27582d 100644
@@ -439,8 +439,7 @@ static int mtk_iommu_v1_create_mapping(struct device *dev,
 	mtk_mapping = data->mapping;
 	if (!mtk_mapping) {
 		/* MTK iommu support 4GB iova address space. */
-		mtk_mapping = arm_iommu_create_mapping(&platform_bus_type,
-						0, 1ULL << 32);
+		mtk_mapping = arm_iommu_create_mapping(dev, 0, 1ULL << 32);
 		if (IS_ERR(mtk_mapping))
 			return PTR_ERR(mtk_mapping);
 
drivers/media/platform/ti/omap3isp/isp.c +1 -1
index 1cda23244c7b..91101ba88ef0 100644
@@ -1965,7 +1965,7 @@ static int isp_attach_iommu(struct isp_device *isp)
 	 * Create the ARM mapping, used by the ARM DMA mapping core to allocate
 	 * VAs. This will allocate a corresponding IOMMU domain.
 	 */
-	mapping = arm_iommu_create_mapping(&platform_bus_type, SZ_1G, SZ_2G);
+	mapping = arm_iommu_create_mapping(isp->dev, SZ_1G, SZ_2G);
 	if (IS_ERR(mapping)) {
 		dev_err(isp->dev, "failed to create ARM IOMMU mapping\n");
 		return PTR_ERR(mapping);
Signature · 1 line
2.34.1
author maintainer
#35 lore raw

PATCH v3 17/21 ARM: dma-mapping: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

Since arm_iommu_create_mapping() now accepts the device, let's replace iommu_domain_alloc() with iommu_paging_domain_alloc() to retire the former.

---
 arch/arm/mm/dma-mapping.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
arch/arm/mm/dma-mapping.c +4 -2
index 52f9c56cc3cb..88c2d68a69c9 100644
@@ -1585,9 +1585,11 @@ arm_iommu_create_mapping(struct device *dev, dma_addr_t base, u64 size)
 
 	spin_lock_init(&mapping->lock);
 
-	mapping->domain = iommu_domain_alloc(dev->bus);
-	if (!mapping->domain)
+	mapping->domain = iommu_paging_domain_alloc(dev);
+	if (IS_ERR(mapping->domain)) {
+		err = PTR_ERR(mapping->domain);
 		goto err4;
+	}
 
 	kref_init(&mapping->kref);
 	return mapping;
Signature · 1 line
2.34.1
Reviewed-by on v3 17/21
#36 lore raw

Re: [PATCH v3 17/21] ARM: dma-mapping: Use iommu_paging_domain_alloc()

Quoted 7 lines · Lu Baolu

Since arm_iommu_create_mapping() now accepts the device, let's replace iommu_domain_alloc() with iommu_paging_domain_alloc() to retire the former.

---
 arch/arm/mm/dma-mapping.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

Jason

author maintainer
#37 lore raw

PATCH v3 18/21 drm/rockchip: Use iommu_paging_domain_alloc()

In reply to Lu Baolu

Commit <421be3ee36a4> ("drm/rockchip: Refactor IOMMU initialisation") has refactored rockchip_drm_init_iommu() to pass a device that the domain is allocated for. Replace iommu_domain_alloc() with iommu_paging_domain_alloc() to retire the former.

---
 drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/gpu/drm/rockchip/rockchip_drm_drv.c +7 -3
index ab55d7132550..52126ffb9280 100644
@@ -103,13 +103,17 @@ static int rockchip_drm_init_iommu(struct drm_device *drm_dev)
 	struct rockchip_drm_private *private = drm_dev->dev_private;
 	struct iommu_domain_geometry *geometry;
 	u64 start, end;
+	int ret;
 
 	if (IS_ERR_OR_NULL(private->iommu_dev))
 		return 0;
 
-	private->domain = iommu_domain_alloc(private->iommu_dev->bus);
-	if (!private->domain)
-		return -ENOMEM;
+	private->domain = iommu_paging_domain_alloc(private->iommu_dev);
+	if (IS_ERR(private->domain)) {
+		ret = PTR_ERR(private->domain);
+		private->domain = NULL;
+		return ret;
+	}
 
 	geometry = &private->domain->geometry;
 	start = geometry->aperture_start;
Signature · 1 line
2.34.1
Reviewed-by on v3 18/21
#38 lore raw

Re: [PATCH v3 18/21] drm/rockchip: Use iommu_paging_domain_alloc()

Quoted 9 lines · Lu Baolu

Commit <421be3ee36a4> ("drm/rockchip: Refactor IOMMU initialisation") has refactored rockchip_drm_init_iommu() to pass a device that the domain is allocated for. Replace iommu_domain_alloc() with iommu_paging_domain_alloc() to retire the former.

---
 drivers/gpu/drm/rockchip/rockchip_drm_drv.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

Jason

author maintainer
#39 lore raw

PATCH v3 19/21 drm/tegra: Remove call to iommu_domain_alloc()

In reply to Lu Baolu

Commit <17de3f5fdd35> ("iommu: Retire bus ops") removes iommu ops from the bus structure. The iommu subsystem no longer relies on bus for operations. So iommu_domain_alloc() interface is no longer relevant.

Normally, iommu_paging_domain_alloc() could be a replacement for iommu_domain_alloc() if the caller has the right device for IOMMU API use. Unfortunately, this is not the case for this driver.

Iterate the devices on the platform bus and find a suitable device whose device DMA is translated by an IOMMU. Then use this device to allocate an iommu domain. The iommu subsystem prevents domains allocated by one iommu driver from being attached to devices managed by any different iommu driver.

---
 drivers/gpu/drm/tegra/drm.c | 34 +++++++++++++++++++++++++---------
 1 file changed, 25 insertions(+), 9 deletions(-)
drivers/gpu/drm/tegra/drm.c +25 -9
index 03d1c76aec2d..ee391f859992 100644
@@ -1133,6 +1133,17 @@ static bool host1x_drm_wants_iommu(struct host1x_device *dev)
 	return domain != NULL;
 }
 
+static int iommu_mapped_device(struct device *dev, void *data)
+{
+	struct device **iommu_dev = data;
+
+	if (!device_iommu_mapped(dev))
+		return 0;
+
+	*iommu_dev = dev;
+	return 1;
+}
+
 static int host1x_drm_probe(struct host1x_device *dev)
 {
 	struct tegra_drm *tegra;
@@ -1149,16 +1160,21 @@ static int host1x_drm_probe(struct host1x_device *dev)
 		goto put;
 	}
 
-	if (host1x_drm_wants_iommu(dev) && iommu_present(&platform_bus_type)) {
-		tegra->domain = iommu_domain_alloc(&platform_bus_type);
-		if (!tegra->domain) {
-			err = -ENOMEM;
-			goto free;
+	if (host1x_drm_wants_iommu(dev)) {
+		struct device *iommu_dev = NULL;
+
+		bus_for_each_dev(&platform_bus_type, NULL, &iommu_dev, iommu_mapped_device);
+		if (iommu_dev) {
+			tegra->domain = iommu_paging_domain_alloc(iommu_dev);
+			if (IS_ERR(tegra->domain)) {
+				err = PTR_ERR(tegra->domain);
+				goto free;
+			}
+
+			err = iova_cache_get();
+			if (err < 0)
+				goto domain;
 		}
-
-		err = iova_cache_get();
-		if (err < 0)
-			goto domain;
 	}
 
 	mutex_init(&tegra->clients_lock);
Signature · 1 line
2.34.1
author maintainer
#40 lore raw

PATCH v3 20/21 iommu: Remove iommu_present()

In reply to Lu Baolu

The iommu_present() interface is no longer used in the tree anymore. Remove it to avoid dead code.

---
 include/linux/iommu.h |  6 ------
 drivers/iommu/iommu.c | 25 -------------------------
 2 files changed, 31 deletions(-)
include/linux/iommu.h +0 -6
index 46f1348f1f0a..ba0b27afc256 100644
@@ -776,7 +776,6 @@ static inline void iommu_iotlb_gather_init(struct iommu_iotlb_gather *gather)
 }
 
 extern int bus_iommu_probe(const struct bus_type *bus);
-extern bool iommu_present(const struct bus_type *bus);
 extern bool device_iommu_capable(struct device *dev, enum iommu_cap cap);
 extern bool iommu_group_has_isolated_msi(struct iommu_group *group);
 extern struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus);
@@ -1072,11 +1071,6 @@ struct iommu_iotlb_gather {};
 struct iommu_dirty_bitmap {};
 struct iommu_dirty_ops {};
 
-static inline bool iommu_present(const struct bus_type *bus)
-{
-	return false;
-}
-
 static inline bool device_iommu_capable(struct device *dev, enum iommu_cap cap)
 {
 	return false;
drivers/iommu/iommu.c +0 -25
index e03c71a34347..c7ebdf96e4bc 100644
@@ -1846,31 +1846,6 @@ int bus_iommu_probe(const struct bus_type *bus)
 	return 0;
 }
 
-/**
- * iommu_present() - make platform-specific assumptions about an IOMMU
- * @bus: bus to check
- *
- * Do not use this function. You want device_iommu_mapped() instead.
- *
- * Return: true if some IOMMU is present and aware of devices on the given bus;
- * in general it may not be the only IOMMU, and it may not have anything to do
- * with whatever device you are ultimately interested in.
- */
-bool iommu_present(const struct bus_type *bus)
-{
-	bool ret = false;
-
-	for (int i = 0; i < ARRAY_SIZE(iommu_buses); i++) {
-		if (iommu_buses[i] == bus) {
-			spin_lock(&iommu_device_lock);
-			ret = !list_empty(&iommu_device_list);
-			spin_unlock(&iommu_device_lock);
-		}
-	}
-	return ret;
-}
-EXPORT_SYMBOL_GPL(iommu_present);
-
 /**
  * device_iommu_capable() - check for a general IOMMU capability
  * @dev: device to which the capability would be relevant, if available
Signature · 1 line
2.34.1
Reviewed-by on v3 20/21
#41 lore raw

Re: [PATCH v3 20/21] iommu: Remove iommu_present()

Quoted 8 lines · Lu Baolu

The iommu_present() interface is no longer used in the tree anymore. Remove it to avoid dead code.

---
 include/linux/iommu.h |  6 ------
 drivers/iommu/iommu.c | 25 -------------------------
 2 files changed, 31 deletions(-)

Jason

author maintainer
#42 lore raw

PATCH v3 21/21 iommu: Remove iommu_domain_alloc()

In reply to Lu Baolu

The iommu_domain_alloc() interface is no longer used in the tree anymore. Remove it to avoid dead code.

There is increasing demand for supporting multiple IOMMU drivers, and this is the last bus-based thing standing in the way of that.

---
 include/linux/iommu.h |  6 ------
 drivers/iommu/iommu.c | 36 ------------------------------------
 2 files changed, 42 deletions(-)
include/linux/iommu.h +0 -6
index ba0b27afc256..bff6e7e81fa3 100644
@@ -778,7 +778,6 @@ static inline void iommu_iotlb_gather_init(struct iommu_iotlb_gather *gather)
 extern int bus_iommu_probe(const struct bus_type *bus);
 extern bool device_iommu_capable(struct device *dev, enum iommu_cap cap);
 extern bool iommu_group_has_isolated_msi(struct iommu_group *group);
-extern struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus);
 struct iommu_domain *iommu_paging_domain_alloc(struct device *dev);
 extern void iommu_domain_free(struct iommu_domain *domain);
 extern int iommu_attach_device(struct iommu_domain *domain,
@@ -1076,11 +1075,6 @@ static inline bool device_iommu_capable(struct device *dev, enum iommu_cap cap)
 	return false;
 }
 
-static inline struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus)
-{
-	return NULL;
-}
-
 static inline struct iommu_domain *iommu_paging_domain_alloc(struct device *dev)
 {
 	return ERR_PTR(-ENODEV);
drivers/iommu/iommu.c +0 -36
index c7ebdf96e4bc..8212fed27e18 100644
@@ -1975,42 +1975,6 @@ __iommu_group_domain_alloc(struct iommu_group *group, unsigned int type)
 	return __iommu_domain_alloc(dev_iommu_ops(dev), dev, type);
 }
 
-static int __iommu_domain_alloc_dev(struct device *dev, void *data)
-{
-	const struct iommu_ops **ops = data;
-
-	if (!dev_has_iommu(dev))
-		return 0;
-
-	if (WARN_ONCE(*ops && *ops != dev_iommu_ops(dev),
-		      "Multiple IOMMU drivers present for bus %s, which the public IOMMU API can't fully support yet. You will still need to disable one or more for this to work, sorry!\n",
-		      dev_bus_name(dev)))
-		return -EBUSY;
-
-	*ops = dev_iommu_ops(dev);
-	return 0;
-}
-
-/*
- * The iommu ops in bus has been retired. Do not use this interface in
- * new drivers.
- */
-struct iommu_domain *iommu_domain_alloc(const struct bus_type *bus)
-{
-	const struct iommu_ops *ops = NULL;
-	int err = bus_for_each_dev(bus, NULL, &ops, __iommu_domain_alloc_dev);
-	struct iommu_domain *domain;
-
-	if (err || !ops)
-		return NULL;
-
-	domain = __iommu_domain_alloc(ops, NULL, IOMMU_DOMAIN_UNMANAGED);
-	if (IS_ERR(domain))
-		return NULL;
-	return domain;
-}
-EXPORT_SYMBOL_GPL(iommu_domain_alloc);
-
 /**
  * iommu_paging_domain_alloc() - Allocate a paging domain
  * @dev: device for which the domain is allocated
Signature · 1 line
2.34.1
Reviewed-by on v3 21/21
#43 lore raw

Re: [PATCH v3 21/21] iommu: Remove iommu_domain_alloc()

Quoted 11 lines · Lu Baolu

The iommu_domain_alloc() interface is no longer used in the tree anymore. Remove it to avoid dead code.

There is increasing demand for supporting multiple IOMMU drivers, and this is the last bus-based thing standing in the way of that.

---
 include/linux/iommu.h |  6 ------
 drivers/iommu/iommu.c | 36 ------------------------------------
 2 files changed, 42 deletions(-)

Jason

#44 lore raw
In reply to Lu Baolu
Quoted 8 lines · Lu Baolu

Lu Baolu (20):
iommu: Add iommu_paging_domain_alloc() interface
iommufd: Use iommu_paging_domain_alloc() vfio/type1: Use iommu_paging_domain_alloc() drm/msm: Use iommu_paging_domain_alloc()
wifi: ath10k: Use iommu_paging_domain_alloc()
wifi: ath11k: Use iommu_paging_domain_alloc() RDMA/usnic: Use iommu_paging_domain_alloc()

Joerg, can you pick at least the above patches for this cycle? Hopefully more will get tags and can go as well.

If you want to try for more tags on the rest I suggest Lu can send patches to the subsystem trees and hope the maintainers merge them. At v6.11-rc5 we declare the remainder unmaintained and you take anything that hasn't been merged so we can finish the iommu core work.

Thanks,
Jason

maintainer
#45 lore raw
In reply to Lu Baolu
Quoted 8 lines · Lu Baolu

The IOMMU subsystem has undergone some changes, including the removal of iommu_ops from the bus structure. Consequently, the existing domain allocation interface, which relies on a bus type argument, is no longer relevant:

struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)

[...]

Applied a few of these to iommu (iommufd/paging-domain-alloc), thanks!

[01/21] iommu: Add iommu_paging_domain_alloc() interface

        https://git.kernel.org/iommu/c/a27bf2743cb8
[02/21] iommufd: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/26a581606fab
[03/21] vfio/type1: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/60ffc4501722
[04/21] vhost-vdpa: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/9c159f6de1ae
[05/21] drm/msm: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/45acf35af200

[10/21] wifi: ath10k: Use iommu_paging_domain_alloc()

        https://git.kernel.org/iommu/c/d5b7485588df
[11/21] wifi: ath11k: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/ef50d41fbf1c

[14/21] RDMA/usnic: Use iommu_paging_domain_alloc()

        https://git.kernel.org/iommu/c/3b10f25704be
[15/21] iommu/vt-d: Add helper to allocate paging domain
        https://git.kernel.org/iommu/c/9e9ba576c259

Cheers,

Signature · 5 lines
Will

https://fixes.arm64.dev
https://next.arm64.dev
https://will.arm64.dev
author maintainer
#46 lore raw
Quoted 23 lines · Will Deacon
Quoted 8 lines · Lu Baolu

The IOMMU subsystem has undergone some changes, including the removal of iommu_ops from the bus structure. Consequently, the existing domain allocation interface, which relies on a bus type argument, is no longer relevant:

struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)

[...]

Applied a few of these to iommu (iommufd/paging-domain-alloc), thanks!

[01/21] iommu: Add iommu_paging_domain_alloc() interface

         https://git.kernel.org/iommu/c/a27bf2743cb8
[02/21] iommufd: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/26a581606fab
[03/21] vfio/type1: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/60ffc4501722
[04/21] vhost-vdpa: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/9c159f6de1ae
[05/21] drm/msm: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/45acf35af200

[10/21] wifi: ath10k: Use iommu_paging_domain_alloc()

Quoted 8 lines
         https://git.kernel.org/iommu/c/d5b7485588df
[11/21] wifi: ath11k: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/ef50d41fbf1c

[14/21] RDMA/usnic: Use iommu_paging_domain_alloc()

         https://git.kernel.org/iommu/c/3b10f25704be
[15/21] iommu/vt-d: Add helper to allocate paging domain
         https://git.kernel.org/iommu/c/9e9ba576c259

Will, the patch [15/21] has already been included in my VT-d update pull request. I have also addressed Yi's comment in that patch. So can you please remove it from this branch?

Thanks,
baolu

maintainer
#47 lore raw
Quoted 28 lines · Baolu Lu
Quoted 27 lines · Will Deacon
Quoted 8 lines · Lu Baolu

The IOMMU subsystem has undergone some changes, including the removal of iommu_ops from the bus structure. Consequently, the existing domain allocation interface, which relies on a bus type argument, is no longer relevant:

struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)

[...]

Applied a few of these to iommu (iommufd/paging-domain-alloc), thanks!

[01/21] iommu: Add iommu_paging_domain_alloc() interface

         https://git.kernel.org/iommu/c/a27bf2743cb8
[02/21] iommufd: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/26a581606fab
[03/21] vfio/type1: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/60ffc4501722
[04/21] vhost-vdpa: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/9c159f6de1ae
[05/21] drm/msm: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/45acf35af200

[10/21] wifi: ath10k: Use iommu_paging_domain_alloc()

         https://git.kernel.org/iommu/c/d5b7485588df
[11/21] wifi: ath11k: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/ef50d41fbf1c
Quoted 8 lines
Quoted 4 lines

[14/21] RDMA/usnic: Use iommu_paging_domain_alloc()

         https://git.kernel.org/iommu/c/3b10f25704be
[15/21] iommu/vt-d: Add helper to allocate paging domain
         https://git.kernel.org/iommu/c/9e9ba576c259

Will, the patch [15/21] has already been included in my VT-d update pull request. I have also addressed Yi's comment in that patch. So can you please remove it from this branch?

Heh, was just about to say the same thing as I noticed when I was re-creating 'next'. Yes, I'll drop it now. Since it's on the end, the other SHAs will remain the same.

Sorry about that,

Will

#48 lore raw
In reply to Will Deacon
Quoted 24 lines · Will Deacon
Quoted 8 lines · Lu Baolu

The IOMMU subsystem has undergone some changes, including the removal of iommu_ops from the bus structure. Consequently, the existing domain allocation interface, which relies on a bus type argument, is no longer relevant:

struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)

[...]

Applied a few of these to iommu (iommufd/paging-domain-alloc), thanks!

[01/21] iommu: Add iommu_paging_domain_alloc() interface

        https://git.kernel.org/iommu/c/a27bf2743cb8
[02/21] iommufd: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/26a581606fab
[03/21] vfio/type1: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/60ffc4501722
[04/21] vhost-vdpa: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/9c159f6de1ae
[05/21] drm/msm: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/45acf35af200

[10/21] wifi: ath10k: Use iommu_paging_domain_alloc()

Quoted 8 lines
        https://git.kernel.org/iommu/c/d5b7485588df
[11/21] wifi: ath11k: Use iommu_paging_domain_alloc()
        https://git.kernel.org/iommu/c/ef50d41fbf1c

[14/21] RDMA/usnic: Use iommu_paging_domain_alloc()

        https://git.kernel.org/iommu/c/3b10f25704be
[15/21] iommu/vt-d: Add helper to allocate paging domain
        https://git.kernel.org/iommu/c/9e9ba576c259

Great, Lu can you please split the remaining by subsystem and try to get them to go through subsystem trees? Joerg can take the leftovers at rc6/7 or something like that so we can finish this.

I think this is enough that Intel and AMD could now assert that dev is non-null?

Jason

author maintainer
#49 lore raw
Quoted 28 lines · Jason Gunthorpe
Quoted 27 lines · Will Deacon
Quoted 8 lines · Lu Baolu

The IOMMU subsystem has undergone some changes, including the removal of iommu_ops from the bus structure. Consequently, the existing domain allocation interface, which relies on a bus type argument, is no longer relevant:

struct iommu_domain *iommu_domain_alloc(struct bus_type *bus)

[...]

Applied a few of these to iommu (iommufd/paging-domain-alloc), thanks!

[01/21] iommu: Add iommu_paging_domain_alloc() interface

         https://git.kernel.org/iommu/c/a27bf2743cb8
[02/21] iommufd: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/26a581606fab
[03/21] vfio/type1: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/60ffc4501722
[04/21] vhost-vdpa: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/9c159f6de1ae
[05/21] drm/msm: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/45acf35af200

[10/21] wifi: ath10k: Use iommu_paging_domain_alloc()

         https://git.kernel.org/iommu/c/d5b7485588df
[11/21] wifi: ath11k: Use iommu_paging_domain_alloc()
         https://git.kernel.org/iommu/c/ef50d41fbf1c
Quoted 7 lines
Quoted 4 lines

[14/21] RDMA/usnic: Use iommu_paging_domain_alloc()

         https://git.kernel.org/iommu/c/3b10f25704be
[15/21] iommu/vt-d: Add helper to allocate paging domain
         https://git.kernel.org/iommu/c/9e9ba576c259

Great, Lu can you please split the remaining by subsystem and try to get them to go through subsystem trees? Joerg can take the leftovers at rc6/7 or something like that so we can finish this.

Yes, sure. I guess I need to wait until 6.11-rc1 is out. It's a bit late in the cycle for the subsystem trees to accept brand new changes.

Quoted 2 lines

I think this is enough that Intel and AMD could now assert that dev is non-null?

Yes.

Thanks,
baolu